From: Keith Seitz <keiths@redhat.com>
To: Daniel Knezevic <daniel.knezevic@htecgroup.com>,
"gdb-patches@sourceware.org" <gdb-patches@sourceware.org>
Cc: Tom Tromey <tom@tromey.com>, Pedro Alves <pedro@palves.net>,
Simon Marchi <simark@simark.ca>
Subject: Re: [PATCH v3] gdb/symtab.c Fix completion of template class members
Date: Tue, 3 Mar 2026 10:19:51 -0800 [thread overview]
Message-ID: <44171ae9-cd60-4c6d-950d-30276cd489eb@redhat.com> (raw)
In-Reply-To: <20260303103633.256157-1-daniel.knezevic@htecgroup.com>
Hi,
On 3/3/26 2:36 AM, Daniel Knezevic wrote:
> This is v3 of:
> https://inbox.sourceware.org/gdb-patches/20251113095559.404539-1-daniel.knezevic@htecgroup.com/raw
>
> Do not break the symbol if it contains '<' or '>' characters.
>
> This patch fixes formatting of the code and adds more tests for testing
> completion of nested template classes.
Thank you for the quick update. I just have a few trivial nits. There is
no need to submit a v4 solely for these.
> diff --git a/gdb/testsuite/gdb.cp/cpcompletion.cc b/gdb/testsuite/gdb.cp/cpcompletion.cc
> index b85f168928c..defbd75ad34 100644
> --- a/gdb/testsuite/gdb.cp/cpcompletion.cc
> +++ b/gdb/testsuite/gdb.cp/cpcompletion.cc
> @@ -74,6 +74,28 @@ struct baz
> S s;
> };
>
> +/* The Outer class is used to test completion of nested template classes*/
Please end the comment with a period and two spaces. [That follows our
coding style guidelines.]
> +template <typename T>
> +struct Outer
> +{
> + T value;
> +
> + template <typename U>
> + struct Inner
> + {
> + U inner_value;
> + T outer_ref;
> + };
> +
> + Inner<int> get_int_inner () const
> + {> + Inner<int> i;
> + i.inner_value = 0;
> + i.outer_ref = value;
> + return i;
> + }
> +};
> +
> int main ()
> {
> baz<int> obj (2.3, 0.1);
> @@ -84,5 +106,9 @@ int main ()
> Foo foo1;
> foo1.set_foo (42); // Set breakpoint here.
> a.get(); // Prevent compiler from throwing 'a' away.
> + Outer<double> outer;
> + outer.value = 4.2;
> + auto inner = outer.get_int_inner();
Need a space between function/method name and parentheses.
> + Outer<double>::Inner<long> long_inner;
> return 0;
> }
> diff --git a/gdb/testsuite/gdb.cp/cpcompletion.exp b/gdb/testsuite/gdb.cp/cpcompletion.exp
> index 1fb9a6bf01d..9ae17026f7d 100644
> --- a/gdb/testsuite/gdb.cp/cpcompletion.exp
> +++ b/gdb/testsuite/gdb.cp/cpcompletion.exp
> @@ -137,3 +137,39 @@ with_test_prefix "expression with namespace" {
> }
>
> test_gdb_complete_unique "break baz(int" "break baz(int, double)"
> +
> +# Test completion with a template class.
> +test_gdb_complete_unique "p baz<int>::" "p baz<int>::baz<double>(int, double)"
> +
> +# Test completion of nested template classes
> +with_test_prefix "nested template classes" {
> + # Unlike in linespecs, tab- and complete-command completion work a
> + # bit differently when completing around the scope operator. The
> + # matches in the tab-completion case only show the part of the
> + # symbol after the scope, since ':' is a word break character.
> +
> + set tab_completion_list {
> + "Inner<int>"
> + "Inner<long>"
> + "get_int_inner() const"
> + }
> + test_gdb_complete_tab_multiple "ptype Outer<double>:" ":" $tab_completion_list
> + test_gdb_complete_tab_multiple "ptype Outer<double>::" "" $tab_completion_list
> +
> + # OTOH, the complete command must show the whole command, with
> + # qualified symbol displayed as entered by the user.
> + set cmd_completion_list {
> + "Outer<double>::Inner<int>"
> + "Outer<double>::Inner<long>"
> + "Outer<double>::get_int_inner() const"
> + }
> + test_gdb_complete_cmd_multiple "ptype " "Outer<double>:" $cmd_completion_list
> + test_gdb_complete_cmd_multiple "ptype " "Outer<double>::" $cmd_completion_list
> + test_gdb_complete_cmd_multiple "ptype " "Outer<double>::Inn" {
> + "Outer<double>::Inner<int>"
> + "Outer<double>::Inner<long>"
> + }
> +
> + # Add a disambiguating character and we get a unique completion.
> + test_gdb_complete_unique "ptype Outer<double>::g" "ptype Outer<double>::get_int_inner() const"
This last line looks a bit long. Suggest adding '\' between the
two arguments, moving the second argument to the next line.
> +}
Otherwise, this LGTM.
Reviewed-By: Keith Seitz <keiths@redhat.com>
Thank you!
Keith
next prev parent reply other threads:[~2026-03-03 18:20 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-03-03 10:36 Daniel Knezevic
2026-03-03 18:19 ` Keith Seitz [this message]
2026-04-20 12:03 ` Daniel Knezevic
2026-04-27 15:28 ` Keith Seitz
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=44171ae9-cd60-4c6d-950d-30276cd489eb@redhat.com \
--to=keiths@redhat.com \
--cc=daniel.knezevic@htecgroup.com \
--cc=gdb-patches@sourceware.org \
--cc=pedro@palves.net \
--cc=simark@simark.ca \
--cc=tom@tromey.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox